The proposed research has as its long-tern goal the development of a quantitative theory of discrimination learning and choice behavior. The experiments are done in a choice situation in which one of a pair of stimuli is presented on each trial, and one response is rewarded if it occurs in the presence of one of these stimuli, and a second response is rewarded if it occurs in the presence of the other stimulus. A large number of the parameters that are known to affect learning in other situations will be systematically studied in this one. The proposed theoretical analysis is based on statistical decision theory.